Professional Documents
Culture Documents
3, May 2016
ABSTRACT
Steganography is the art of hiding information in ways that prevent the detection of hidden messages. This
paper presentsa new method which randomly selects position in MP3 file to hide a text secret messageby
using Least Significant Bit (LSB) technique. The text secret message isused in start and ends locations a
unique signature or key.The methodology focuses to embed one bit, two bits, three bitsor four bits from
secret message into MP3 file by using LSB techniques. The evaluation and performancemethods are based
on robustness (BER and correlation), Imperceptibility (PSNR) and hiding capacity (Ratio between Sizes of
text message and MP3 Cover) indicators.The experimental results show the new method is more security.
Moreover the contribution of this paper is the provision of a robustness-based classification of LSB
steganography models depending on their occurrence in the embedding position.
KEYWORDS
Steganography, LSB, mp3 data set, hiding a message
1. INTRODUCTION
Steganography is the art and science of hiding information by embedding messages within others,
seemingly harmless messages. Steganography means covered writing in Greek. As the goal of
steganography is to hide the presence of a message and create a covert channel, it can be seen as
the complement of cryptography, whose goal is to hide the content of a message [1].
Steganography basically aims at hiding communication between two parties from the attackers
[3]. Steganography operates by embedding a secret message which might be a copyright mark, or
a covert communication, or a serial number in a cover message such as a video film, an audio
recording, or computer code in such a way that it cannot be accessed by any wrong person during
data exchange.
The three types of Steganography include the first, Pure Steganography where there is no need for
the key. , Second Secret Key steganography and lastly, Public Key Steganography is based on the
concepts of public key cryptography. Public key steganography uses a public key and a private
key to secure the communication between the parties [20].
Steganography technique used in the data hiding process must have important properties in order
to secure data successfully. Some of these properties include robustness, imperceptibility and
capacity. These properties are explained below. Robustness means resistance to blind, nontargeted modifications, or common image operations [6]. Imperceptibility is typically required for
DOI: 10.5121/ijcnc.2016.8305
67
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
secure covert communication. For example, if a steganography method uses the noise component
of digital images to embed a secret message, it should do so while not making statistically
significant changes to the noise in the carrier.Capacity or (Data Rate) refers to the amount of
information that can be hidden relative to the size of the cover message [21].
Capacity
Imperceptibility
Robustness
2. STENOGRAPHIC METHODS
This section analyses the steganography which help understand the topic in a new perspective
Steganography methods can be classified mainly into six categories, although in some cases exact
classification is not possible [2].
Substitution methods substitute redundant parts of a cover with a secret message (spatial
domain). A number of methods exist for hiding information in various media. These
methods range from LSB coding also known as bit plane or noise insertion
toolsmanipulation of image or compression algorithms to modification of image properties
such as luminance. Basic substitution systems try to encode secret information by
substituting insignificant parts of the cover by secret message bits; the receiver can extract
the information if he has knowledge of the positions where secret information has been
embedded. Since only minor modifications are made in the embedding process, the sender
assumes that they will not be noticed by a passive attacker.[14]
Transform domain techniques embed secret information in a transform space of the signal
(frequency domain): It has been noted early in the development of stenographic systems
that embedding information in the frequency domain of a signal can be much more robust
than embedding rules operating in the time domain. Most robust stenographic systems
known today actually operate in some sort of transform domain.[15]
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
the data, and a synchronized reception with the code at the receiver is used for dispreading
and subsequent data recovery.[16]
Distortion techniques store information by signal distortion and measure the deviation from
the original cover in the decoding step.
Cover generation methods encode information in the way a cover for secret communication
is created.
Size of file
(WAVE)
Size under
320kbps
MP3
Size under
256kbps
MP3
Size under
192kbps
MP3
Size under
128kbps
MP3
Size under
96kbps
MP3
Classical
14.7
6.67
5.33
2.66
Jazz
Country
16.2
18.7
7.34
8.48
5.87
6.78
4.4
5.08
2.93
3.39
2.2
2.54
R&B
19.4
8.81
7.05
5.29
3.52
2.64
Rap
20.1
9.14
7.31
5.48
3.65
2.74
Reggae
20.1
9.14
7.31
5.48
3.65
2.74
69
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
Pop
20.2
9.16
7.33
5.49
3.66
2.75
Rock
23
10.4
8.35
6.26
4.17
3.13
Blues
11.8
10.7
8.59
6.44
4.29
3.22
Hip-hop
27.5
12.4
9.98
7.48
4.99
3.74
Dance
31.3
14.2
11.3
8.53
5.68
4.26
Metal
32.6
14.8
11.8
8.88
5.92
4.44
Name of Text
100.txt
200.txt
400.txt
800.txt
1Mb.txt
2Mb.txt
Size
100 KB
200 KB
400 KB
800 KB
1 MB
2 MB
Size byte
102,590 bytes
205,180 bytes
410,364 bytes
819,932 bytes
1,049,704 bytes
2,099,408 bytes
4. PROPOSED ALGORITHM
To make stego MP3 file by reading an MP3 audio file and text message file then, embed the text
file inside the audio MP3 to generate stego MP3 that contains a specific message.
70
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
(1)
(2)
Where, r is the number of samples in the MP3 filerb*cb is the size of text message file
deg is the number insertion bitsrand is a function that generates random number from 0 to
1.This will generates a random starting
starting location of embedding where it ensures that, the end of
insertion will be located inside the MP3 file.
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
matrix of two-dimension format, as shown in the example bellow of the Hello word. The row
indicates the character, and the columns represent the binary code for the specific character.
>>dec2bin (ans)
ans =
1001000
1100101
1101100
1101100
1101111
We used function to converts the decimal numbers to binary as the following equation where Bdi
is the binary digit index, and Ddi+1 is the decimal digit division result. And the rem is the
division remainder.
Bdi=rem (Ddi+1/2)
(3)
The results will be a two dimensional matrix with size of Rx8 where the 8 is the number of bits
for ASCII character conversion to binary, and the R is the number of characters in the text file. R
is counting not only alphabetic characters, but also any ASCII symbol including the space and
carriage return.
(4)
After normalized sample audio processing is to convert the normalized audio samples to binary
format. The conversion function is the same that has been used for the text file conversion that is
described before.
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
10101010101010
01010101010101
10101010101010, 01010101010101
01010101010101, 10101010101010
The same signature is being used for start and end of embedding. So, the massage is being
bounded within the same signature that indicates start and end of the message, in addition to
number of insertion bits. The embedding is being done in terms of insertion. The insertion is a
method that is simply removes a bit or number of bits from the carrier data and inserts new bit or
bits from the message data.
The figure 2 bellow shows how to insert one bit, and two bits from the message to the carrier. The
insertion is being done in terms of least significant bit. From (least significant), its noticeable
effect of the actual digital value is negligible. So, the insertion will not highly affect the resulted
audio data with respect to the person who hears the MP3 audio file.
There are four insertion scenarios depending on the number of bits those are inserted within the
carrier, according to the developed system. Single bit insertion, two bits insertion, three bits
insertion, and four bits insertion is possible and those are selected with the input arguments of the
developed program. The figure above shows how to insert single bit, and two bits only. The
described process above is continued until all text data bits inserted within the digital carrier data
that represents the audio MP3 data.
Figure 3. the insert 1 bit, and 2 bits from the message to the carrier
Once the all-binary samples of text file were inserted and the process is completed, an inverse
process will be performed in the post-processing phase. The resulted digital array that represented
the stego MP3 file will be converted to decimal again in an inverse process of that described
above. Where above a decimal to binary conversion is being accomplished, but now, binary to
decimal will be performed. The following equation illustrates how to convert the binary data to
decimal.
73
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
Dd=bi*2i
(5)
Where Dd is the decimal digit that is resulted after conversion, bi is the ith binary bit value, i is the
index of the binary bit. The i has values from 0 to 22. Hence, the maximum normalized decimal
number is 2x106, thats why the maximum number of i is 22.The resulted decimal data is
normalized according to the normalization process that is described above. Thus, denormalization is required to get the analog audio format again. The following equation is being
used for de-normalization.
Ai= (AiN*106)-1
(6)
The latest step is finally to convert the stego audio data to an MP3 format file. The same toolbox
for MP3 format handling is also used; it contains an MP3 write function. It takes the number of
bits and MP3 encoding format to generate formal MP3 file. The same MP3 parameters those are
gotten when read of the original MP3 file are used to write the new stego MP3 file.
In figure 4The following flowchart that explain process of the model that start to read the text
and MP3 file and then Normalize MP3 file and convert the text file to ASCII format, and then
Embeddedbits within the MP3 file, after finish that we reverse the process to restores the MP3
file as normal MP3.
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
5. EXPERIMENTAL RESULTS
The Peak Signal-to-Noise Ratio (PSNR) is the ratio between a signal's maximum power and the
power of the signal's noise. Engineers commonly use the PSNR to measure the quality of
reconstructed signals that have been compressed. Signals can have a wide dynamic range, so
PSNR is usually expressed in decibels as show Comparisonsbetween embedded different bits in
figure 7. In statistics, the Mean Squared Error (MSE) of an estimator is one of many ways to
quantify the difference between values implied by an estimator and the true values of the quantity
being estimated. MSE is a risk function, corresponding to the expected value of the squared error
loss or quadratic loss. MSE measures the average of the squares of the "errors." The error is the
amount by which the value implied by the estimator differs from the quantity to be estimated, as
show Comparisonsbetween embedded different bits in figure 6. [19].
Table 3. PSNR and MSE values of ten audio files at 200 kb text message s
Name of genre
Classical
Jazz
Country
R&B
Rap
Reggae
Pop
Rock
Blues
Dance
Hip-Hop
Metal
PSNR
75.4490
63.0154
63.1514
63.7896
58.7419
60.8632
65.4141
65.4814
66.9008
64.1514
66.2314
63.4814
MSE
0.0019
0.0325
0.0315
0.0272
0.0869
0.0533
0.0187
0.0184
0.0133
0.0175
0.0168
0.0135
No. of Embedding
7714000
8494000
9805000
10195000
10573000
10571000
10596000
12068000
12423000
13486000
13926000
14499000
In Error! Reference source not found.5 shows the signal structure of the audio file Rap.MP3 at
before embedding. By using MP3 audio with 128kbps and the size of MP3 3.65 MB (3,836,190
bytes) and the genre Rap. Also use 200 KB the secret Message.
Figure 5. the signal level comparisons between a MP3 carrier file before and after the LSB with size 3.65
MB Text 200 KB
75
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
Table 4. PSNR and MSE values of ten audio files at 200 kb text message when embedded 1 bit and
embedded 2 bits
Name of genre
embedded 1 bit
PSNR
Classical
78.1266
0.0010
No. of
Embedding
7714
77.6796
0.0011
No. of
Embedding
7714
Jazz
Country
63.4224
63.5267
0.0296
0.0289
8494
9805
63.4224
63.5267
0.0296
0.0289
8494
9805
R&B
Rap
64.1845
59.1376
0.0248
0.0793
10194
10572
64.1845
59.1376
0.0248
0.0793
10194
10572
Reggae
Pop
61.2687
66.1688
0.0486
0.0157
10571
10595
61.2687
65.7593
0.0486
0.0173
10571
10595
65.8772
67.3400
0.0168
0.0120
12067
12422
65.8772
67.3400
0.0168
0.0120
12067
12422
Dance
Hip-Hop
67.0223
60.1148
0.0129
0.0633
16443
14423
65.2804
58.8159
0.0193
0.0854
16443
14423
Metal
68.1388
0.0100
17114
66.8664
0.0134
17114
Rock
Blues
MSE
embedded 2 bits
PSNR
MSE
Table 5 PSNR and MSE values of ten audio files at 200 kb text message when embedded 3 bits and 4 bits
Name of genre
embedded 3 bits
PSNR
MSE
embedded 4 bits
Classical
77.6796
0.0014
No. of
Embedding
7714
Jazz
Country
R&B
Rap
Reggae
Pop
Rock
Blues
Dance
Hip-Hop
Metal
63.4224
63.5267
64.1845
59.1376
61.2687
65.7593
65.8772
67.3400
65.2804
58.8159
66.8664
0.0267
0.0262
0.0225
0.0719
0.0439
0.0142
0.0152
0.0108
0.0193
0.0854
0.0134
8494
9805
10194
10572
10571
10595
12067
12422
16443
14423
17114
PSNR
MSE
77.2308
0.0012
No. of
Embedding
7714
64.3006
64.3789
65.0483
59.9930
62.1421
67.0086
66.7580
67.9554
65.2817
58.8168
66.8681
0.0242
0.0237
0.0203
0.0651
0.0397
0.0129
0.0137
0.0110
0.0193
0.0854
0.0134
8494
9805
10194
10572
10571
10595
12067
12422
16443
14423
17114
76
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
0.12
0.1
0.08
MSE 1 bit
0.06
MSE 2 bits
MSE 3bits
0.04
MSE 4 bits
0.02
0
Figure 6. Comparisons
omparisons between different results for Mean Squared Error (MSE)
80
70
60
50
PSNR 1 bit
40
PSNR 2 bits
30
PSNR 3bits
20
PSNR 4 bits
10
0
Figure 7. Comparisons between different results for Mean Squared Error (MSE)
6.CONCLUSIONS
This paper has explored and reviewed MP3 audio steganography, particularly with respect to
MP3 files after compression. LSB in time domain has been developed to use randomly position
from cover file to hide the secret message by using 1,
1 2, 3 and 4 bits. The new Model aims at
meeting the three most important audio steganography requirements, which are imperceptibility,
capacity, and robustness. Any technique tries to enhance the capacity or robustness should
77
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
preserve imperceptibility. A new method is increased the capacity and robustness as well as
improved the imperceptibility. In this paper, we concentrate model that has been built, achieved
hiding the data in Audio file, by keeping the accuracy of the audio file high, even though, one
mange to discover the secret message, still extracting the message is challenging.
6. ACKNOWLEDGEMENTS
This Project was funded by the Deanship of Scientific Research (DSR) at King Abdulaziz
University, Jeddah, under grant no. (G-481-611-37). The authors, therefore, acknowledge with
thanks DSR for technical and financial support
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Katzenbeisser S., Peticotas F., Information Hiding Techniques For Steganography And Digital
Watermarking, Artech House Inc.2000
Stefan Katzenbeiser& Fabien A.P.Petitcolas(1999), Information Hiding Techniques for
Steganography and Digital Watermarking, Artech House, Computer Security series, Boston, London
Kivanc M., Information Hiding Codes And Their Applications To Images And Audio, Phd. Thesis,
University Of Illinois At Urbana-Champaign, 2002.
Cacciaguerra S., Ferretti S., Data Hiding: Steganography And Copyright Marking, Department Of
Computer Science, University Of Bologna, Italy, Url: Http://Www.Cs.Unibo.It/~Scacciag/HomeFile/Teach/Datahiding.Pdf.
Dunbar B., A Detailed Look AtSteganographic Techniques And Their Use In An Open-Systems
Environment, Sans Institute 2002, Url: Http://Www.Securitydoc.Com/Library/1272.
Fridrich, J. (2010). Steganography in Digital Media Principles, Algorithms, and Applications.
Cambridge University Press: UK.
Bender W., Gruhl D., Morimoto N., Lu A., "Techniques For Data Hiding , Ibm System Journal, Vol.
35, No. 3&4,1996,Url: Http://Isj.Www.Media.Mit.Edu/Isj/Sectiona/313.Pdf.
P.K. Singh, H. Singh, And K. Saroha, A Survey On Steganography In Audio, Audio, 2009.
M. Wakiyama, Y. Hidaka, And K. Nozaki, An Audio Steganography By A Low-Bit Coding Method
With Wave Files, 2010 Sixth International Conference On Intelligent Information Hiding And
Multimedia Signal Processing, Oct. 2010, Pp. 530-533
Kekre, H. B., Athawale, a, Rao, B. S., &Athawale, U. (2010). Increasing the Capacity of the Cover
Audio Signal by Using Multiple LSBs for Information Hiding. 2010 3rd International Conference on
Emerging Trends in Engineering and Technology, 196-201. IEEE. doi:10.1109/ICETET.2010.118
Zamani, M., Taherdoost, H., Manaf, A. A., Ahmad, R. B., &Zeki, A. M. (2009). Robust Audio
Steganography via Genetic Algorithm. Soft Computing, 0-4.
Zamani, M., Manaf, A. A., Ahmad, R. B., Zeki, A. M., & Abdullah, S. (2009). A Genetic-AlgorithmBased Approach for Audio Steganography. Engineering and+ Technology, 360-363.
Bhowal, K., Pal, a J., Tomar, G. S., &Sarkar, P. P. (2010). Audio Steganography Using GA. 2010
International Conference on Computational Intelligence and Communication Networks, 449-453.
Ieee. doi:10.1109/CICN.2010.91
Cox, I., et al., "A Secure, Robust Watermark for Multimedia," in Information Hiding: First
International Workshop, Proceedings, vol. 1174 of Lecture Notes in Computer Science, Springer,
1996, pp. 185206
Koch, E., and J. Zhao, "Towards Robust and Hidden Image Copyright Labeling," in IEEE Workshop
on Nonlinear Signal and Image Processing, Jun. 1995, pp. 452455.
Tirkel, A. Z., G. A. Rankin, and R. van Schyndel, "Electronic Watermark," in Digital Image
Computing, Technology and ApplicationsDICTA 93, Macquarie University, 1993, pp. 666673
Rckert, C., R. Szczepanowski, et al.(2005) "Complete genome sequence of the actinobacterium<
i>Actinoplanesfriuliensis</i> HAG 010964, producer of the lipopeptide antibiotic friulimycin."
Journal of biotechnology 178: 41-42.
Chanu, Y. J., K. M. Singh, et al. "Image steganography and steganalysis: A survey." International
Journal of Computer Applications 52(2).2012
Atoumet al., A New Method for Audio Steganography Using Message Integrity .Journal of
Convergence Information Technology(JCIT),Volume8, Number14, September 2013
78
International Journal of Computer Networks & Communications (IJCNC) Vol.8, No.3, May 2016
[20] Atoum, M. S., Ibrahim, S., Sulong,G. and Ahmed, A. (2012). MP3 Steganography: Review. Journal
of Computer Science issues, 9(6).
[21] Atoum, M. S. (2015). A Comparative Study of Combination with Different LSB Techniques in MP3
Steganography. In Information Science and Applications (pp. 551-560). Springer Berlin Heidelberg
[22] Atoum, M. S. (2015, August). New MP3 Steganography Data Set. In IT Convergence and Security
(ICITCS), 2015 5th International Conference on (pp. 1-7). IEEE.
AUTHORS
Mr.Alaaalarood He still continues his Ph.D. in UniversityTechnology Malaysia (UTM)
Faculty of computingand Information Technology. He is Lecturer of king abed alaziz
university, Department of Computing and Information Technology.His research interests
are Information Security,steganalysis, steganography, Artificial Intelligence ANN, and
Computer Graphics.
Prof. Azizah Abdul Manaf, Professor of Computer Science, Deputy Dean Academic
Advanced Informatics School (UTM AIS) UniversitiTeknologi Malaysia (UTM) her
research interests are Image Processing, Multimedia Security, Computer Forensics.
79